NVIDIA Unleashes The First Jetson AGX Orin Module

Again in March, NVIDIA launched Jetson Orin, the following-era of their ARM single-board computers supposed for edge computing apps. The new platform promised to provide “server-class AI performance” on a board smaller ample to set up in a robotic or IoT unit, with even the least expensive tier of Orin modules presenting approximately double the functionality of the past Jetson Xavier modules. Sadly, there was a little bit of a catch — at the time, Orin was only readily available in progress package variety.

But currently, NVIDIA has introduced the rapid availability of the Jetson AGX Orin 32GB creation module for $999 USD. This is primarily the mid-array providing of the Orin line, which makes releasing it to start with a sensible more than enough preference. Users who need to have the top-end overall performance of the 64GB variant will have to wait until finally November, but there’s nevertheless no tough launch day for the lesser NX Orin SO-DIMM modules.

That’s a little bit of a letdown for individuals like us, due to the fact the two SO-DIMM modules are probably the most pleasing for hackers and makers. At $399 and $599, their pricing tends to make them much far more palatable for the personal experimenter, though their lesser dimensions and much more common interface must make them much easier to carry out into Diy builds. Though the Jetson Nano is still an unbeatable discount for those searching to dip their toes into the CUDA waters, we could unquestionably see individuals investing in the much much more powerful NX Orin boards for additional sophisticated tasks.

Though the AGX Orin modules may well be a bit steep for the common tinkerer, their availability is nevertheless some thing to be enthusiastic about. Many thanks to the typical JetPack SDK framework shared by the Jetson household of boards, applications produced for these increased-stop modules will mostly continue being suitable throughout the whole product or service line. Sure, the less costly and older Jetson boards will operate them slower, but as significantly as device discovering and AI apps go, they’ll continue to operate circles close to something like the Raspberry Pi.